1 Hayashi 289-325 - law.ua.edu

1 HAYASHI 289-325 (DO NOT DELETE) 11/14/2017 8:29 PM

289

A THEORY OF FACTS AND CIRCUMSTANCES

Andrew T. Hayashi*

ABSTRACT .................................................................................................. 290 INTRODUCTION .......................................................................................... 290 I. GENERAL STATEMENT OF THE PROBLEM .............................................. 293

A. How Private Information Matters in Tax Law ........................ 296 B. The Costs of Deceit .................................................................. 298

II. A PRINCIPLE FOR CHOOSING FACTS AND CIRCUMSTANCES ................. 300 A. Screening and Tagging ............................................................ 304 B. Backflips with a Twist .............................................................. 305 C. Application: Hobby Losses ...................................................... 307

III. A PRINCIPLED STANDARD FOR FACTS AND CIRCUMSTANCES

TESTS .............................................................................................. 313 A. Choosing Between Two Games ............................................... 315 B. A Principled Standard ............................................................. 316 C. Rules and Standards as Games ............................................... 322

CONCLUSION .............................................................................................. 324

* Associate Professor of Law, University of Virginia School of Law. Thanks to Ben Doherty, Steven Keithley, and Andrew Bae for excellent research assistance. Thanks also to Dave Anderson, Michael Doran, Barbara Fried, Brian Galle, Yehonatan Givati, Mitchell Kane, Leandra Lederman, Francine Lipman, Ruth Mason, Greg Mitchell, Susie Morse, Kelly Mulholland, Henry Ordower, Gregg Polsky, Diane Ring, Deborah Schenk, Kirk Stark, Paul Stephan, Ethan Yale, George Yin, Larry Zelenak, seminar participants at the Duke University School of Law, St. Louis University School of Law, The University of Nevada, Las Vegas Boyd School of Law, the 2016 University of Virginia Invitational Tax Conference, the 2016 National Tax Association Annual Meeting, and the 2017 Stanford/Harvard/Yale Junior Faculty Forum.


290 Alabama Law Review [Vol. 69:2:289

ABSTRACT

The legal consequences of an action often depend on information that only the actor knows. This information is typically inferred from the observable “facts and circumstances” attending the actor’s conduct, which creates a seemingly unresolvable tension in legal design. On the one hand, these unstructured inquiries give free rein to the factfinder’s judgment about which facts justify an inference about the hidden information. On the other hand, specifying the facts that will be used to draw that inference would provide a roadmap for actors to adjust their conduct strategically to manipulate the factfinder’s conclusions. I argue that this tension can be resolved by applying insights from the economics literature on asymmetric information. These insights help answer both the substantive question of which facts and circumstances should be taken into account and the procedural question of whether they should be specified by the legislature or left to the courts.

INTRODUCTION

Legal consequences often turn on things that cannot be observed. This is true in criminal law, where the actor’s state of mind is an element of most crimes,1 and in contract law and labor law, which impose obligations to perform and negotiate in “good faith.”2 But it is even true in areas of the law such as federal income taxation, despite persistent skepticism among commentators that taxes should depend on anything other than publicly observable actions.3 For example, the tax consequences of a transaction may be disregarded if the taxpayer does not have a substantial nontax

1. Morissette v. United States, 342 U.S. 246, 252 (1952) (discussing the development of a mens rea element in most crimes).

2. See Continental Ins. Co. v. NLRB, 495 F.2d 44, 48 (2d Cir. 1974) (“[P]arties are obligated to do more than merely go through the formalities of negotiation.”); Murphy v. Allstate Ins. Co., 553 P.2d 584, 586 (Cal. 1976) (“This court has observed that ‘[i]n every contract there is an implied covenant of good faith and fair dealing’ . . . .” (alteration in original) (citation omitted)); 1 N. PETER LAREAU, LABOR AND EMPLOYMENT LAW § 12.05, LexisNexis (2017); 30 SAMUEL WILLISTON & RICHARD A. LORD, WILLISTON ON CONTRACTS § 77:10 (4th ed.), Westlaw WILLSTN-CN (database updated May 2017). On the obligation to bargain in good faith, see 29 U.S.C. § 158(d) (2012) and NLRB v. Boss Mfg. Co., 118 F.2d 187, 189 (7th Cir. 1941) (good faith means negotiating “with an open and fair mind and sincerely endeavor[ing] to overcome obstacles or difficulties”), all discussed in Samuel W. Buell, Good Faith and Law Evasion, 58 UCLA L. REV. 611, 633–34 (2011).

3. See, e.g., Edwin S. Cohen, Tax Avoidance Purpose as a Statutory Text in Tax Legislation, 9 PROC. ANN. TUL. TAX INST. 229, 257 (1960) (“[T]he tax structure should be satisfied without our seeking to gauge the extent of [the taxpayer’s] consciousness in a hazardous effort to probe his state of mind.”); Leandra Lederman, W(h)ither Economic Substance?, 95 IOWA L. REV. 389, 433 (2010) (“[T]ax motivation should make no difference as to whether or not the claimed tax result is upheld . . . .”)


2017] A Theory of Facts and Circumstances 291

purpose for that transaction,4 and expenses are deductible if incurred for the purpose of generating income but not if those expenses are personal in nature.5 When legal consequences turn on things that cannot be observed, factfinders often look to the facts and circumstances surrounding the conduct to draw inferences about these unobservable factors. But there is a tension at the heart of facts and circumstances inquiries.

On the one hand, the open-ended nature of these inquiries creates uncertainty about their outcomes, which ultimately depend only on the factfinder’s judgment and experience about how the observable facts relate to the hidden factor. Factfinders are generally given little guidance about which facts permit a reliable inference to these hidden factors, and factfinders rarely provide reasons for relying on the facts that they do. As a result, people cannot know which facts a judge or jury will regard as relevant or how those facts will be interpreted, and so they have no way of reliably communicating to the factfinder whether they possess the hidden factor on which legal consequences depend.

It would be a natural reaction to this uncertainty to provide more detail in the applicable law. Specifying the facts used to draw inferences about hidden factors would solve the problem of uncertainty; however, in doing so, it would create another problem. If the facts that create a favorable inference about a hidden factor are publicized in advance, they will provide a roadmap for well-advised individuals to create those very facts to induce factfinders to draw the inference those individuals want. As a result, the factfinder will be unable to distinguish between those who genuinely possess the hidden factor and those who mimic them.

In this Article, I propose a two-part solution to this tension. The first part addresses the substantive question of which facts and circumstances should be considered by factfinders. I propose a “screening principle” for choosing the facts and circumstances to be used in drawing an inference about hidden information. By placing restrictions on the kinds of facts to be considered, the principle structures the facts and circumstances inquiry and reduces the possibility that idiosyncratic differences between factfinders will result in arbitrary differences in legal outcomes. The principle resolves the tension between uncertainty and strategic mimicry by selecting for consideration only those facts that are much costlier for mimics to emulate than for parties who genuinely possess the hidden factors. Thus, even if mimics are on notice about which kinds of facts will be used to draw an inference about a hidden factor, the cost of mimicry will be too great to make it worthwhile.

4. I.R.C. § 7701(o) (2012). 5. I.R.C. §§ 212, 262 (2012).



The second part of the solution addresses when the screening principle should be used to specify the facts relevant to a facts and circumstances determination. I consider whether the principle should be used to specify the relevant facts and circumstances in the applicable statute or regulations, so that people can act in full knowledge of which facts will be considered, or whether the principle should be used by courts to interpret conduct after it has occurred. This is a choice between a rule and a standard, and I provide an economic analysis that is based on, and extends, Louis Kaplow’s foundational work on this choice.6

Professor Kaplow’s analysis focuses on laws that depend on observable actions,7 and I show that extending the analysis to laws that depend on hidden factors introduces another evaluative dimension to the choice between a rule and a standard. In these cases, individuals and lawmakers are playing a “game” in which the factfinder tries to distinguish between people who have some hidden factor and others who only pretend that they do. The choice between a rule and a standard in this setting is a choice between two different games with potentially different outcomes. Adopting a rule creates what the game theory literature calls a “screening” game, whereas a standard establishes a “signaling” game. After taking this new dimension into account, I argue that laws that depend on hidden factors are best implemented as a particular kind of standard that I call a “principled standard.” I also argue that the signaling games created by principled standards raise additional concerns and discuss how to manage them.

This Article is about how to structure facts and circumstances inquiries, and I assume that it is worthwhile, in at least some circumstances, to conduct such inquiries rather than adopt objective rules that do not depend on intentions or other hidden factors but may nevertheless provide rough justice. In many cases, rough justice is good enough, and the economics framework that I use to generate this proposal also helps identify when facts and circumstances tests are not worth the candle. For example, when the screening principle can effectively distinguish individuals by their hidden characteristics, then facts and circumstances tests become cheaper to administer and more attractive. Conversely, when it is not possible to identify factors that satisfy the principle, then my analysis provides another reason to favor per se rules. Thus, my analysis both helps identify when facts and circumstances inquiries should be avoided altogether and provides a structure for applying them when they are necessary.

Throughout this Article I use federal income tax law to illustrate the tension inherent in facts and circumstances tests and to show how the tension can be resolved using my proposal. Income tax law considers

6. Louis Kaplow, Rules Versus Standards: An Economic Analysis, 42 DUKE L.J. 557 (1992). 7. See id. at 557.



unobservable taxpayer information to be normatively relevant in many contexts, and the problem of strategic behavior by taxpayers is especially acute. The centrality of these two considerations in tax law makes it the ideal setting to illustrate the application of my proposal.8 Part I explains the role hidden factors play in tax law. In Part II, I explain the screening principle and how it emerges from the economics literature on asymmetric information, and I describe how the principle can address the important tax distinction between business and personal expenses. In Part III, I address the question of whether laws that depend on private information should be promulgated as rules or standards. I conclude that such laws are best implemented as principled standards.

I. GENERAL STATEMENT OF THE PROBLEM

Facts and circumstances inquiries are used to solve a common problem: the legal consequences of an activity depend on some information that is known only by the actor. Because the factfinder cannot observe these crucial facts, it must infer them from what it can observe. For example, whether a transfer of property is a gift for federal income tax purposes depends on the intentions of the transferor, turning on whether it was motivated by “detached and disinterested generosity.”9 How is a factfinder to know what the transferor’s intentions were? Courts look to the observable facts and circumstances surrounding the transfer to determine the transferor’s intent. But which facts should they look to? Which facts permit a valid inference to the transferor’s intentions? Courts do not say.10 In Commissioner v. Duberstein, the Court concluded that “[d]ecision of the issue presented in these cases must be based ultimately on the application of the fact-finding tribunal’s experience with the mainsprings of human conduct to the totality of the facts of each case.”11 Concurring in part and

8. Another feature of tax disputes that makes tax law a particularly relevant setting for my analysis is the explicit permissibility of forum shopping. Taxpayers who are unable to resolve a matter within IRS administrative proceedings have a choice about whether to litigate the dispute in Tax Court, federal district court, or the Court of Federal Claims. David J. Herzig, Justice for All: Reimagining the Internal Revenue Service, 33 VA. TAX REV. 1, 4 (2013). Not only might the choice of forum affect the substantive law applicable to the taxpayer’s dispute, but it might also affect the way judges sitting in those courts tend to interpret the facts.

9. Comm’r v. Duberstein, 363 U.S. 278, 285 (1960). 10. See Evangelista v. Comm’r, 629 F.2d 1218, 1222 (7th Cir. 1980) (“Eschewing an opportunity

to establish an easy to apply test, the Court stated that the determination of whether a transfer is a gift depends on an objective inquiry of all the circumstances . . . .”); Poyner v. Comm’r, 301 F.2d 287, 292 (4th Cir. 1962) (“The Court limited itself to summarizing earlier decisions as to which particular dominant motivations, when adequately supported by the evidence, result in income treatment, and which result in gift treatment. An enumeration of the criteria, by which the trier of fact shall determine in every type of case what that dominant reason is, was deemed inadvisable, if not futile.”).

11. Duberstein, 363 U.S. at 289.



dissenting in part, Justice Frankfurter articulates well one of the primary concerns with such an open-ended approach:

Varying conceptions regarding the “mainsprings of human conduct” are derived from a variety of experiences or assumptions about the nature of man, and “experience with human affairs,” is not only diverse but also often drastically conflicting. What the Court now does sets fact-finding bodies to sail on an illimitable ocean of individual beliefs and experiences. This can hardly fail to invite, if indeed not encourage, too individualized diversities in the administration of the income tax law.12

To take another example from federal income tax law, consider that expenses incurred “for [the] production of income” are deductible whereas “personal” expenses are not.13 How is a factfinder to determine the purpose behind incurring the expense and whether the taxpayer received some personal (i.e., nonpecuniary) benefit as a result? Taxpayers have been left with Justice Cardozo’s memorable, but dispiriting, conclusion that “[o]ne struggles in vain for any verbal formula that will supply a ready touchstone. The standard set up by the statute is not a rule of law; it is rather a way of life. Life in all its fullness must supply the answer to the riddle.”14

One problem that taxpayers face in “sail[ing] on an illimitable ocean of individual beliefs and experiences”15 of factfinders is that they get different outcomes depending on how their judge or jury thinks that gift givers behave, or what sorts of purchases and expenditures generate personal pleasure and which ones do not. The uncertainty about which facts a court will look to is compounded by the uncertainty about how those facts will be weighed against each other if they suggest opposite conclusions.16

So why not specify, in the applicable statute or regulations, which facts will be used to draw an inference about individuals’ private information? One reason, of course, is the wide range of circumstances that facts and circumstances tests are meant to address. There may simply be too many facts that have too different evidentiary values in too many different situations to specify them all in advance. The second reason not to specify

12. Id. at 297 (Frankfurter, J., concurring in part and dissenting in part). 13. I.R.C. §§ 262, 212, 162 (2012). 14. Welch v. Helvering, 290 U.S. 111, 115 (1933). Although Justice Cardozo purported to be

analyzing the difference between a current and capital business expense and the analysis turned on the question of whether the expense was “ordinary” within the meaning of the statute, the Court was also worried about providing a deduction for expenses that are personal in nature. Id. at 114.

15. Duberstein, 363 U.S. at 297 (Frankfurter, J., concurring in part and dissenting in part). 16. See Exacto Spring Corp. v. Comm’r, 196 F.3d 833, 835 (7th Cir. 1999) (the multifactor test

for whether a salary qualifies as “reasonable compensation” is “nondirective. No indication is given of how the factors are to be weighed in the event they don’t all line up on one side.”).



the facts is that doing so would provide a strategic advantage to the regulated parties. Once Congress or the Treasury Department specifies the facts from which factfinders must draw an inference favorable to the taxpayer, then all taxpayers have an incentive to produce those facts.17 These efforts to persuade the tax authority about the taxpayer’s motivations are socially wasteful, and it is undesirable to reward mimics who engage in this kind of deception. Giving taxpayers a roadmap of how to manipulate the factfinder’s inference would undermine the law’s ability to distinguish between the mimics and the people who genuinely have the hidden factor.

For example, suppose that the factfinder believes that gifts always prompt the recipient to write a thank-you note, but that transfers that have some other purpose, such as to nurture a commercial relationship, do not. Gifts are not taxed to the recipient.18 The recipients of true gifts may write a thank-you note out of a sense of gratitude, but the recipients of commercial inducements may do so too, merely to ensure that the factfinder believes that they received a “gift.” Moreover, recipients of veritable gifts who would ordinarily prefer to express their gratitude in some other way may also switch to note writing to ensure that the factfinder is certain of the giver’s intentions.

A more important example of this phenomenon is written tax opinions, which are sought by taxpayers from lawyers and accountants. A taxpayer who wants to comply with her tax obligations in good faith will often want to seek the advice of a tax professional to learn the legal basis for the position she takes on her return with respect to that transaction. Historically, taxpayers who obtained such an opinion could use it as evidence of that good faith, which provides a defense against the imposition of civil penalties for the underpayment of tax on account of a reporting position that was ultimately held to be invalid.19 Perhaps unsurprisingly, taxpayers with no interest in tax compliance began to obtain written opinions from less scrupulous lawyers and accountants for the sole purpose of avoiding penalties with respect to transactions that they knew would not withstand scrutiny.20

Thus, when taxpayers are told what the IRS considers to be the visible trappings of someone entitled to a favorable tax result, they will adopt those trappings. But, when taxpayers do not know what the factfinder will interpret as evidence of the hidden factor, the problem can become even worse. Taxpayers may engage in all sorts of peripheral and socially

17. The feedback effect resembles the issues described in Joshua B. Fischman, The Circular Logic of Actavis, 66 AM. U. L. REV. 91 (2016).

18. I.R.C. § 102 (2012) (gifts excluded from income). 19. I.R.C. § 6664(c) (2012). 20. Dennis J. Ventry, Jr. & Bradley T. Borden, Probability, Professionalism, and Protecting

Taxpayers, 68 TAX LAW. 83, 110 (2014).



wasteful activities because of this uncertainty, just to increase the likelihood that they will receive the tax treatment that they want.

In the rest of the Article, I use examples from federal income tax law to illustrate how insights from game theory can both reduce the unpredictability of facts and circumstances tests and ensure that regulated parties do not manipulate factfinders into drawing the wrong inference. Tax law uses intentions and other private information to assign tax consequences and also must deal with aggressively strategic behavior by taxpayers to misrepresent those intentions, making it an ideal place to explore my proposal.21

A. How Private Information Matters in Tax Law

Nearly fifty years ago, Walter Blum observed that “[u]nder our federal income tax many of the substantive rules for classifying actions have long appeared to call for an inquiry into somebody’s state of mind,”22 such as their motive, intent, or purpose for taking the action.23 But why are a taxpayer’s purposes relevant to the tax treatment of their transactions? There are many places in the tax law that a taxpayer’s intentions are relevant, but most of them can be gathered under two themes: intentions that are relevant to measuring income, and intentions that are used to police aggressive tax-avoidance schemes.24 In each case there has been a lively debate about whether the tax law should take intentions into account, but I do not take sides in those debates. My aim is to show how facts and circumstances inquiries into intent and other hidden factors should be structured, when it has already been decided that such inquiries are necessary.

The first reason that the tax law incorporates private information is because the definition of income depends on it. Most tax scholars accept, at

21. There are other contexts in which facts and circumstances tests are used, but where one or both of these considerations is absent. For example, sometimes the totality of the circumstances is used to arrive at a legal conclusion, rather than an inference. An example of this usage is the determination of whether a particular financial instrument is debt or equity for federal income tax purposes. Courts will look to a variety of facts to arrive at a conclusion about whether the instrument is, in fact, debt. Estate of Mixon v. United States, 464 F.2d 394, 402 (5th Cir. 1972). Whether the instrument is debt does not depend on some information known only to the holder of the financial instrument.

22. Walter J. Blum, Motive, Intent, and Purpose in Federal Income Taxation, 34 U. CHI. L. REV. 485, 485 (1967).

23. According to Blum, motive is about the reasons why the action was taken, purpose is what was hoped to be accomplished, and intent is about whether a consequence was within the expectation of the person acting. Id. at 486–87. It matters little, in Blum’s view, whether we refer to motive, intent, or purpose. Id. at 487.

24. Daniel N. Shaviro, In Defense of Requiring Back-Flips, 26 VA. TAX REV. 815, 815 (2007). In Shaviro’s view, tax shelters are not malum in se; instead, “the issue is one of malum prohibitum, or acts that are wrong simply in a compliance sense once they have been identified as legally impermissible or ineffective.” Id. at 816.



least as a starting point, Henry Simons’s definition of income as a person’s consumption plus her change in wealth over a period of time.25 Accordingly, a net income tax must distinguish between expenses incurred for the production of income and those that yield consumption.26 What is consumption? In many cases it can be difficult to know whether an individual has incurred an expense to purchase a consumption benefit or to earn income. As Simons himself noted: “[H]ere one finds inescapable the unwelcome criterion of intention . . . Given items will represent business expense in one instance and merely consumption in another.”27

The second reason that a taxpayer’s intentions play an important role in tax law and administration is because they are used to police behavior that is undertaken to reduce tax liability inappropriately. Specifically, transactions that lack a business purpose altogether, or which have as a principal purpose the avoidance of federal income tax, may be disallowed.28

25. HENRY C. SIMONS, PERSONAL INCOME TAXATION: THE DEFINITION OF INCOME AS A

PROBLEM OF FISCAL POLICY 50 (1938) (income is “the algebraic sum of (1) the market value of rights exercised in consumption and (2) the change in the value of the store of property rights between the beginning and end of the period in question”).

26. The test of whether an activity is entered into for profit is the taxpayer’s subjective intent. Metz v. Comm’r, 2015 T.C.M. (CCH) 1248 (2015); Skeen v. Comm’r, 864 F.2d 93, 94 (1989). In truth, many expenses are motivated by both business and pleasure. In such cases there is not even an answer, in principle, to the question of what the taxpayer’s (one) purpose is.

27. HENRY C. SIMONS, PERSONAL INCOME TAXATION: THE DEFINITION OF INCOME AS A

PROBLEM OF FISCAL POLICY 54 (1938). Numerous sections of the Internal Revenue Code address the blurry line between personal and income-producing expenses. See, e.g., I.R.C. § 162 (2012) (deductions for trade or business expenses); I.R.C. § 212 (2012) (expenses incurred for the production of income); I.R.C. § 262 (2012) (the nondeductibility of personal expenses); I.R.C. § 183 (2012) (rules limiting the deductibility of losses incurred in an activity not engaged in for profit).

28. The tax consequences of a transaction are upheld only if the taxpayer “has a substantial purpose (apart from Federal income tax effects) for entering into such transaction.” I.R.C. § 7701(o) (2012). Intent-based antiavoidance rules also appear in the Treasury Regulations. For example, consider the partnership antiabuse regulation in Treas. Reg. § 1.701-2, which states that subchapter K (pertaining to partnerships) requires a business purpose for any transaction or series of transactions entered into by a partnership. Treas. Reg. § 1.701-2(a)(1) (1995) (“The partnership must be bona fide and each partnership transaction or series of related transactions (individually or collectively, the transaction) must be entered into for a substantial business purpose.”). I do not consider here the advisability or efficiency of antiavoidance doctrines in general. Professor Weisbach makes a compelling case that such doctrines can be an important part of the tax system as a way of broadening the base and reducing the elasticity of taxable income. David A. Weisbach, An Economic Analysis of Anti-Tax-Avoidance Doctrines, 4 AM. L. & ECON. REV. 88 (2002). Such intent-based tests are controversial, both because they seem to raise seemingly intractable line-drawing problems between good tax-motivated behavior and bad tax-motivated behavior, and because of problems of administration and fairness. See, e.g., Alan Gunn, Tax Avoidance, 76 MICH. L. REV. 733, 765 (1978). Gunn argues that, in fact, “the question whether particular conduct was tax-motivated should be irrelevant to the decision whether that conduct should be taxed in a certain way.” Id. at 765. But see David A. Weisbach, Ten Truths About Tax Shelters, 55 TAX L. REV 215, 252 (2002) (“[T]here are good reasons to base anti-avoidance doctrines on motive or intent. There is a difference between somebody engaging in a transaction for purely business reasons that happens to have fantastic tax consequences and somebody entering into the transaction solely to reduce taxes. In the former case, where the taxpayer enters into the transaction for business reasons, there is no economic distortion caused by taxes—while the person pays low taxes,



Intentions are also used to resolve a number of smaller, discrete matters of tax law. For example, gifts are neither deductible by the giver nor included in the recipient’s income.29 Whether a transfer is a gift depends on the motivation of the transferor. If the transfer is made from the “detached and disinterested generosity”30 of the transferor and “out of affection, respect, admiration, charity or like impulses”31 then it is a gift, and “[w]hat controls is the intention with which payment, however voluntary, has been made.”32 Taxpayer intentions also influence whether a series of actions will be aggregated for the purposes of tax analysis. The “step transaction” doctrine is used by courts to determine whether to perform this aggregation.33 Courts have been explicit about the role of intent when applying the step transaction doctrine, doing so when a series of transactions “are really prearranged parts of a single transaction intended from the outset to reach the ultimate result.”34 Once one becomes aware of the importance of intentions in tax law, examples abound.

B. The Costs of Deceit

Intentions matter in tax law. Of course, courts cannot observe intentions, so they must infer those intentions from the facts and circumstances surrounding the taxpayer’s actions.35 As a practical matter, then, laws that depend on intentions ultimately depend on the observable external facts and circumstances from which inferences are drawn.36

behavior is not distorted by this prospect. In the latter case, where the motive is taxes, behavior is distorted, and there are real economic costs. The two cases are different precisely because of mental states.”).

29. I.R.C. § 102 (2012). 30. Comm’r v. Duberstein, 363 U.S. 278, 285 (1960) (quoting Comm’r v. Lo Bue, 351 U.S. 243,

246 (1956)). As Blum noted about gifts, “If differentiations are based on the reason (or reasons) why a transferor gave something of value to the taxpayer . . . the classification cannot avoid inquiring into state of mind.” Blum, supra note 22, at 543. There is a per se rule for gifts from employers to employees in § 102(c) of the Internal Revenue Code. However, Prop. Treas. Reg. § 1.102-1(f)(2), 54 Fed. Reg. 595, 631 (Jan. 9, 1989) allows that an employer can nevertheless make a gift to her employee if that employee is the natural object of her affection.

31. Robertson v. United States, 343 U.S. 711, 714 (1952). 32. Duberstein, 363 U.S. at 286 (quoting Bogardus v. Comm’r, 302 U.S. 34, 45 (1937)

(Brandeis, J., dissenting)). 33. Penrod v. Comm’r, 88 T.C. 1415, 1428 (1987). 34. Id. at 1429. 35. David A. Weisbach, Formalism in the Tax Law, 66 U. CHI. L. REV. 860 (1999); Blum, supra

note 22, at 504 n.44 (“Any state of mind test is likely to develop a set of subordinate rules of thumb which make explicit the external factors that are relevant in ascertaining state of mind. These make possible some consistency of treatment.”).

36. Blum, supra note 22, at 544 (“But while there is a difference between tests based on states of mind and tests based on external factors, we have seen that results under the two types of tests are generally not very far apart. Whenever state of mind is relevant, the most important operational question usually concerns the weight that is to be attached to various external factors.”).



Naturally, it is in the economic interest of taxpayers to disguise or misrepresent their intentions to obtain favorable treatment under a facts and circumstance test, and they are assisted in this endeavor by both professional tax advisors and other, like-minded taxpayers who are facing the same tax issues.37

For example, hobbyists try to pass for someone with a profit motive by maintaining accounting records, creating a website through which they sell their goods and services, and opening separate bank accounts.38 Hobbyists are also often advised to manipulate the timing of expenses and income from their hobby so that they can periodically report positive annual net income; doing this creates a favorable presumption under the hobby-loss rules about whether the activity is engaged in for profit.39 All of these machinations are directed at avoiding limitations on the deductibility of their hobby expenses by emulating people motivated by profit.

Decorating transactions with ancillary activities to disguise the taxpayer’s true intentions is not only a waste of resources, but it also encourages taxpayers to be dishonest. One prominent practitioner worries about intent-based tests for just this reason, arguing that “such doctrines have induced taxpayers and their counsel to take elaborate steps to clothe tax-motivated schemes in the dress of business purpose, thus introducing a

37. Tony Nitti, A Tale of Two Activities: How to Beat the Hobby Loss Rules, FORBES (Oct. 8, 2014, 5:45 PM), http://www.forbes.com/sites/anthonynitti/2014/10/08/a-tale-of-two-activities-how-to-beat-the-hobby-loss-rules/#3d17613a72a9 (providing guidance for “what we should do—and, much more importantly, not do—to beat the hobby loss rules”). Richard J. Kovach, Bright Lines, Facts and Circumstances Tests, and Complexity in Federal Taxation, 46 SYRACUSE L. REV 1287 (1996) (“A construction contractor might wish to avoid employment taxes respecting a crew of carpenters. A competent tax advisor can explain how the contractor and carpenters should alter their behavior to maximize chances for attaining the desired tax result. The objective of such behavior modification would be the creation of a set of facts and circumstances totally in support of the proposition that the carpenters are not employees. Unfortunately, at least some of the proposed changes in behavior are likely to cause great inconvenience to the contractor and carpenters.”). In 2009 the IRS issued a manual for field agents to use when auditing hobby losses. Robert W. Wood, Ten Tips for Deducting Your ‘Hobby’, FORBES (Jan. 8, 2010, 7:20 PM), http://www.forbes.com/2010/01/08/irs-tax-audit-hobby-losses-personal-finance-robert-wood.html. Forbes advised that the new hobby-loss manual for field agents was itself a blueprint for how to arrange one’s affairs to generate a favorable tax conclusion. Tax advisors play a role too, giving advice about how to “help avoid IRS disallowance of losses . . . [or] insure a solid case for the Tax Court.” Jane O. Burns & S. Michael Groomer, Effects of Section 183 on the Business/Hobby Controversy, 58 TAXES 195 (1980). See Wood, supra (“Will the IRS pay for your hobby? The short answer is: No. But the more nuanced answer is: ‘Yes, Uncle Sam will sometimes subsidize your hobby.’ If, that is, you make it into enough of a real business.”). Among the suggestions includes manipulating when expenses are incurred to try to gain the presumption. Wood, supra.

38. See Susan Lee, Hobby Loss: Can Artists Take Losses on Their Tax Returns?, FREELANCETAXATION.COM, https://www.freelancetaxation.com/hobby-loss-can-artists-take-losses-on-their-tax-returns (last visited Feb. 10, 2017); Martin M. Shenkman, Avoid the Hobby Loss Deduction Limits, WORK-FOR-RVERS-AND-CAMPERS.COM, http://www.work-for-rvers-and-campers.com/hobby-loss.html (last visited Feb. 10, 2017).

39. I.R.C. § 183(d) (2012).



significant lack of candor and forthrightness into the relationship between taxpayers and the government.”40

This dishonesty is especially worrisome because it is easy for taxpayers to engage in it without openly acknowledging the artifice. Whereas inventing a business expense whole cloth leaves little room for self-deception that what one is doing is wrong, it is comparatively easy to convince oneself that among the various motives one has for entering into a transaction is some business purpose and that adopting the factors that one knows will persuade a court of favorable tax treatment is simply a cost of revealing to the court something that might be true in any event (or so the taxpayer can tell herself).41

Thus, the problems with facts and circumstances tests are threefold. They invite taxpayers to decorate their activities with the indicia of tax-favored intentions, which is socially wasteful. They cause the intent-based test to be overinclusive and therefore excessively costly, in revenue terms. Finally, they encourage dishonesty.42 My proposal addresses these three concerns.

II. A PRINCIPLE FOR CHOOSING FACTS AND CIRCUMSTANCES

In this Part, I set forth the first aspect of my proposal for structuring facts and circumstances inquiries. I propose that the factfinder use only facts that are much easier to produce for those with the requisite intent. Limiting consideration to only such facts and circumstances will screen out those with the relevant intention or other hidden characteristic by making it much costlier for mimics to emulate people with that characteristic.43

To see how this works, consider the following hypothetical. Andy and Betty are the sole shareholders of Tech Corp., which sells technology

40. Dana L. Trier, Beyond the Smell Test: The Role of Substantive Anti-Avoidance Rules in Addressing the Corporate Tax Shelter Problem, 78 TAXES 62, 62 (2000).

41. As Blum notes: “The process [of inference from external events to intentions] is obviously full of pitfalls. Not only may recordations and recollections be corrupted by possible tax consequences, but the very thoughts that did enter the actor’s consciousness may have been spawned or refined by an awareness of tax considerations.” Blum, supra note 22, at 498.

42. In analyzing intent-based tests, Buell notes other ways that regulated parties may manipulate evidence about their intentions: “Actors may not leave sufficient evidence to permit conclusions ex post about what they were thinking at the time of their conduct. To the extent actors are aware that legal inquiry focuses on mental state, actors may learn to thwart the anti-evasion doctrine itself by taking steps to reduce the availability of such evidence.” Buell, supra note 2, at 664.

43. There are some other technical conditions that we want the screening facts to satisfy. In addition to the negative correlation between the cost of the signal and presence of the characteristic, there should be fine-grained levels of the signal within the appropriate cost range. Michael Spence, Job Market Signaling, 87 Q.J. ECON. 355, 368 (1973). This is a principle in the sense of Ehrlich and Posner, who assert that “[p]roperly understood, ‘principles’ are simply the considerations that are relevant in determining the content of a rule.” Isaac Ehrlich & Richard A. Posner, An Economic Analysis of Legal Rulemaking, 3 J. LEGAL STUD. 257 (1974).



consulting services. The company was formed two years ago when Andy contributed $50,000 in exchange for 50% of the corporation’s common stock and Betty contributed $50,100 for the other 50% of the common stock and all of a special class of preferred stock that entitles her to receive cash distributions, at the discretion of Tech Corp.’s board of directors. Betty is also an employee of Tech Corp., with years of experience and expertise that is highly valued by Tech Corp.’s clients. Shortly after receiving a payment of $500,000 from client X, Tech Corp.’s board declared a dividend of $200,000 to each of Andy and Betty, in respect of their common stock holdings, and a special dividend of $100,000 to Betty, in respect of her preferred stock.

Notwithstanding the “dividend” label, Betty’s status as both a shareholder and an employee of Tech Corp. means that the $100,000 distribution could be either an actual dividend (a return on her invested capital) or compensation for her services. Which of these two treatments is appropriate depends on whether Tech Corp.’s intent in making the distribution was compensatory or not. Determining this intent is a facts and circumstances inquiry, and one factor that courts have used to infer this intent is whether cash distributions are proportional to employee stockholdings.44 This is a good screening factor because shareholders may be willing to tolerate a disproportionate distribution to an employee/shareholder if that distribution serves a compensatory purpose, but not otherwise. Presumably, all shareholders benefit from providing employees with adequate compensation that creates incentives for productive work. A disproportionate dividend, however, will likely just come at the expense of the other shareholders. It is true that the distribution in this case is perhaps not disproportionate to Betty’s proportionate holdings of preferred stock, but it is wildly disproportionate to her invested capital, and one might surmise that the issuance of the preferred stock for only $100 was done solely to persuade a court that the distribution was a dividend.

Using the screening principle to identify the relevant facts and circumstances for inferring private information makes it possible to look to objective factors as a proxy without creating an incentive for taxpayers to adopt wasteful behaviors to mislead the factfinder, thereby solving the key inferential problem in making taxes depend on private information. By focusing on objective factors rather than relying on taxpayers’

44. Kennedy v. Comm’r, 671 F.2d 167, 175 (6th Cir. 1982) (“One factor which is indicative of a distribution of capital rather than compensation is if the payments are in proportion to the employee’s stockholdings. . . . Where compensation paid to officers bears no relation to stockholdings, it tends to show that amounts paid were actually compensation and not disguised dividends.”); see also Paul E. Kummer Realty Co. v. Comm’r, 511 F.2d 313 (8th Cir. 1975); Nor-Cal Adjusters v. Comm’r, 503 F.2d 359 (9th Cir. 1974); Wm. Yuenger Mfg. Co. v. United States, No. 52-C-2643, 1957 WL 10721 (N.D. Ill. Feb. 15, 1957); Berkshire Oil Co. v. Comm’r, 9 T.C. 903 (1947); Soabar Co. v. Comm’r, 7 T.C. 89 (1946).



representations about their intentions, it also avoids the problem that intent-based taxes are viewed as a tax on candor and avoids creating a temptation for taxpayers to be dishonest or misrepresent their intentions; the principle aligns self-interest with an honest revelation of the taxpayers’ intentions.

The screening principle is an efficient way of sorting taxpayers according to their private information, and it can provide the touchstone for which Justice Cardozo struggled in vain in Welch. By being explicit about the principle that facts must satisfy to generate an inference about private information, facts and circumstances inquiries can be made more disciplined and taxpayers will both be able to signal more reliably when they do have the tax-favored intentions and be discouraged from pretending that they have those intentions when they don’t. But why require that courts only use factors that are much less costly for taxpayers with the tax-favored characteristic?

The first reason for focusing on factors that are much less costly for taxpayers with normatively relevant characteristics is efficiency. If we imagine a screening characteristic that is only a little bit less costly for taxpayers with the tax-favored attribute than those without it, taxpayers who want to distinguish themselves from those without the attribute will need to engage in a lot of the activity in order to separate themselves. This is wasteful. If it is not possible to find an activity that is sufficiently differentially costly, depending on the taxpayers’ motivations, then it will be too costly to use facts and circumstances tests to distinguish among them.45

For example, suppose that we observe two individuals who report that they are engaged in the business of being artists.46 Suppose also that each additional hour spent networking with art dealers, traveling and exhibiting their art, and procuring new and better materials is a little less costly for the profit-motivated individual than the one who creates art primarily for recreation. This is plausible because, whatever intrinsic pleasure there may be in these activities, they also generate an increased likelihood of profit from the activity which, by assumption, motivates only the first individual. If the marginal cost of each additional hour spent on these activities is only a little bit less for the profit-motivated individual, then they will need to spend much more time than they would like to distinguish themselves from

45. See John G. Riley, Weak and Strong Signals, 104 SCANDINAVIAN J. ECON. 213 (2002). 46. For expositional purposes, throughout this Article I use examples in which there are two

types of taxpayers. This will tend to be associated with two potential treatments for the transaction that they undertake. For example, deductibility or non-deductibility of an expense, or either gift or compensation treatment for a property transfer. In principle, and in fact, there is likely to be a continuum of types of individuals who vary in the strengths of their motivations by small degrees. Under the appropriate conditions the conclusions in this Article about when it is possible and desirable to separate these individuals, and provide a continuum of tax consequences, would hold.



the recreational artist. On the other hand, if these activities were absolutely loathsome to the recreational artist, then even a couple hours spent on them would be sufficient to distinguish the profit-motivated individual from the recreational artist.

The second reason for focusing on strong screening factors is related to the first. Weak screening factors make it very costly for people with the right motivations to distinguish themselves from people who are trying to mimic them, and if it is too costly to do this, then the people with the right motivations might not undertake the activity at all. Consider again our example of the two artists. If the profit-motivated individual has more attractive alternatives to making art than the recreational artist, then making the deductibility of expenses dependent on a weak screen, and thereby demanding a lot of wasteful activity from the profit-motivated individual to deduct her expenses, will tend to drive those individuals out of the art business. The only individuals who will be left making art are those doing it for recreation.

Thus, there are good theoretical reasons for relying only on factors for which the cost varies significantly with the underlying taxpayer characteristic that we care about. But there is also a pragmatic reason for excluding weak screening factors: it simplifies the analysis for judges and removes unhelpful discretion. One might think that it would be appropriate for courts always to include all factors that are less costly for taxpayers with the tax-favored attribute and simply give greater weight to those that are much less costly. Even if we had not ruled out these weak screens as a theoretical matter, we might worry that courts will struggle with how to appropriately weight each screening factor. Admitting weak screens at the outset creates the risk that those weak screens will be given significant weight by the courts. Some weighting of factors is inevitable, to be sure, but we can simplify the factfinder’s task by asking them only to identify strong screening factors, rather than identify all screening factors and then also weight them appropriately.

The screening principle is not radical and in many areas of the tax law may approximate what courts are already doing.47 I consider this to be a virtue of the principle. The principle both provides a partial explanation for why judges and factfinders choose the facts that they do and is therefore an explanation of what the law is as well as a statement of what it should be. To the extent that the principle states what the law is, it provides some rationalization of what appear on their face to be unstructured and

47. In Part C, I apply the screening principle to evaluate a collection of factors specified by the Treasury Regulations in the rules applicable to hobbies and the evaluation of whether a taxpayer has a profit motive. It appears the two of the factors that satisfy the screening principle may play a meaningful role in determining whether such a profit motive exists.



unconstrained inquiries based on nothing more than the judge’s idiosyncratic views and the “illimitable ocean of individual beliefs and experiences.”48

Screening is deeply embedded in tax law scholarship.49 For this reason, my proposal will be familiar to tax law scholars and is closely connected to existing literature on how the creation of rules that impose different costs on different groups can induce an efficient sorting of people into those receiving favored treatment and those not.50

A. Screening and Tagging

The simple idea that tax treatment should depend on differences between taxpayers that are difficult to mimic is rooted in economics research on screening and tagging.51 The concept of screening is at the heart of “optimal taxation” theory,52 which aims to tax individuals on the productivity of their labor.53 We cannot observe that productivity, so we tax something that is correlated with it: income. The problem, of course, is that individuals who have a high ability to earn income may choose not to work, thereby emulating those who have a low ability to earn. The optimal tax problem is how to tax more heavily those with a higher ability to earn while preventing them from behaving as though they have low ability.54

48. Comm’r v. Duberstein, 363 U.S. 278, 297 (1960) (Frankfurter, J., concurring in part and dissenting in part).

49. Leigh Osofsky, Who’s Naughty and Who’s Nice: Frictions, Screening, and Tax Law Design, 61 BUFF. L. REV. 1057, 1058–59 (2013) (“A large body of literature regarding optimal tax theory addresses screening taxpayers. At base . . . screening mechanisms in the optimal tax context track characteristics indicative of ability and impose greater costs on high ability taxpayers trying to obtain low tax rates or other benefits, in order to target the low rates or benefits more accurately to low ability taxpayers.”).

50. Id. at 1075–77, 1075–77 nn. 61–63. 51. George A. Akerlof, The Market for “Lemons”: Quality Uncertainty and the Market

Mechanism, 84 Q.J. ECON. 488 (1970); Jennifer F. Reinganum, Plea Bargaining and Prosecutorial Discretion, 78 AM. ECON. REV. 713 (1988); Joseph E. Stiglitz, The Theory of “Screening,” Education, and the Distribution of Income, 65 AM. ECON. REV. 283 (1975); William Vickrey, Counterspeculation, Auctions, and Competitive Sealed Tenders, 16 J. FIN. 8 (1961). On the related topic of “ordeals,” see George A. Akerlof, The Economics of “Tagging” as Applied to the Optimal Income Tax, Welfare Programs, and Manpower Planning, 68 AM. ECON. REV. 8 (1978); Albert L. Nichols & Richard J. Zeckhauser, Targeting Transfers through Restrictions on Recipients, 72 AM. ECON. REV. 372 (1982); Tomer Blumkin, Yoram Y. Margalioth & Efraim Sadka, The Role of Stigma in the Design of Welfare Programs 1–23 (CESifo Working Paper Series No. 2305, 2008), https://papers.ssrn.com/abstract=1137843.

52. This approach originated in James Mirrlees’s path-breaking 1971 paper. J. A. Mirrlees, An Exploration in the Theory of Optimum Income Taxation, 38 REV. ECON. STUD. 175 (1971).

53. Id. 54. Screening logic has been used to address a couple of specific problems under the income tax.

Alex Raskolnikov has argued that it is possible to create two regimes for taxpayer compliance: one that resembles current law but with higher penalties for improper reporting, and another with lower penalties but in which taxpayers will have a disfavored position in litigation with the IRS over questionable



The best way to do this, it turns out, is to make taxes dependent on immutable characteristics, things that the taxpayers cannot themselves change (or at least only change at prohibitive expense).55 To the extent that taxes and transfers can be made contingent on characteristics that serve as proxies for earning ability, they can be made more efficient. These characteristics are known as “tags.”56

The logic of screening and tagging as applied to intent-based tests implies that the law must look to behaviors that are correlated with the intention that we wish to favor. These behaviors are known in tax law scholarship as “frictions.”57 The idea that frictions can be an important constraint on tax planning is also a well-understood idea in the tax literature.58

B. Backflips with a Twist

One way of thinking about facts that satisfy the screening principle is that they impose costs, or frictions, on the kind of tax planning that individuals would like to engage in, with the added feature that they impose greater frictions on certain taxpayers than others. Practitioners and tax scholars are well-acquainted with the role that economic risk and other frictions have on the willingness of taxpayers to undertake tax-motivated transactions.59 These costs generally require that taxpayers deviate from their most preferred transaction, and economic substance approaches to

transactions. Alex Raskolnikov, Revealing Choices: Using Taxpayer Choice to Target Tax Enforcement, 109 COLUM. L. REV. 689 (2009). Raskolnikov distinguishes between “separating” and “targeting” features of the two regimes, with the first intended to cause different types of taxpayers to sort between the regimes, and the second designed to facilitate compliance. Larry Zelenak has noted that that we should be concerned about the fairness of subjecting certain taxpayers to a tilted playing field solely to discourage other taxpayers from choosing that regime. Lawrence Zelenak, Tax Enforcement for Gamers: High Penalties or Strict Disclosure Rules?, 109 COLUM. L. REV. (SIDEBAR) 55 (2009). Emily Satterthwaite argues that tax elections can serve as screening devices. Emily Satterthwaite, Tax Elections as Screens, 42 QUEENS L. J. 63 (2016).

55. Akerlof, supra note 51. 56. Id. One example of tagging in tax law is the enhanced standard deduction for individuals with

visual impairment. I.R.C. § 63(f)(2) (2012). More generally, however, the federal income tax does not assign tax liabilities on the basis of immutable characteristics such as race, gender, and height, for reasons that likely have to do with fundamental notions of fairness embodied in the Equal Protection Clause of the federal Constitution.

57. David M Schizer, Frictions as a Constraint on Tax Planning, 101 COLUM. L. REV. 1312, 1315 (2001).

58. Id. 59. As David Schizer has observed, frictions can arise in many ways, including differences in

accounting and regulatory treatment, that discourage taxpayers from switching for tax reasons between economically equivalent transactions. Id. For this reason, Schizer argues that “[s]tudying frictions thus should become a priority for legal commentators.” Id. at 1317.



deterring tax planning are focused on such frictions.60 From an efficiency perspective, the desirability of such frictions depends on two considerations: the importance of deterring tax planning and the effectiveness of the frictions in deterring the planning.61 The nature of the friction, on this account, is unimportant. As Professor Shaviro memorably puts it:

[O]ne might as well condition favorable tax consequences on whether the taxpayer’s chief financial officer can execute [twenty] back-somersaults in the IRS National Office at midnight on April Fool’s Day, if such a requirement turns out to achieve a better ratio of successful deterrence to inducing wasteful effort in meeting requirements that are pointless in themselves.62

Of course, frictions do not affect all taxpayers equally. Leigh Osofsky has advanced our understanding of these effects and argued for the use of frictions as screening devices for tax planning.63 Her criteria for evaluating frictions are:

First, frictions must impose greater costs on tax planners than non-planners. Second, as to tax planners, the friction must deter tax planning, rather than causing it to continue in a more wasteful fashion. Third, the benefits from the first two steps (from increased efficiency and lower tax liability for non-planners) must outweigh costs that taxpayers bear as a result of the friction.64

As an example of a good screen, Osofsky offers the wash-sales rules, which she argues “principally screen[] among taxpayers with different motivations for selling built-in loss stock.”65 The wash-sale rules perform this function because taxpayers that tend to sell and repurchase the same stock in a short period of time are more likely to be tax-motivated than

60. Shaviro says that: “If we did not use economic substance tests to challenge the reality of the cubbyholes taxpayers try to exploit, we would in effect have created a regime of pure electivity to claim whatever losses and ignore whatever gains one likes. The frictions imposed by an economic substance rule burden taxpayer electivity. If tax rules are sufficiently well-designed, they can reduce overall deadweight loss even though they induce particular taxpayers to waste even more resources.” Shaviro, supra note 24, at 818.

61. Daniel N. Shaviro, In Defense of Requiring Back-Flips, 26 VA. TAX REV. 815, 818–19

(2006). 62. Id. (alteration in original) (quoting Daniel N. Shaviro, Economic Substance, Corporate Tax

Shelters, and the Compaq Case, 88 TAX NOTES 221, 223 (2000)). 63. Osofsky, supra note 49. 64. Osofsky, supra note 49, at 1059. 65. Id. at 1094.



not.66 The screening principle I have set forth selects facts that satisfy the criteria for good frictions advanced by Professor Osofsky.

C. Application: Hobby Losses

In 2007, the Treasury Inspector General for Tax Administration published a report finding that 1.5 million taxpayers with significant income from other sources reported only losses on their Schedule C in the years 2002–2005.67 These losses, which may have saved these taxpayers as much as $2.8 billion in taxes in tax year 2005,68 likely arose from the expenses associated with hobbies that taxpayers misrepresented as businesses.69

In this Part, I discuss how the screening principle can be used to tailor the facts and circumstances test for hobby losses. Distinguishing between hobbies and activities engaged in for profit can be difficult, particularly because in the early stages of a business a taxpayer may be working during her free time, generating significant losses, and experimenting with ways of conducting the business that may not pan out.70 Taxpayers have historically taken aggressive positions to exploit this difficulty.71

The hobby-loss rules are also becoming more important with the increasingly unstable nature of employment and the popularity of activities that people use to supplement their income but that also have hobby-like characteristics.72 For example, arts and crafts sold through websites like Etsy and multi-level marketing schemes are frequent targets of the hobby loss rules.73 The following post from a multi-level marketing online discussion forum for a product called “Beachbody” illustrates the issues:

66. Id. 67. TREASURY INSPECTOR GEN. FOR TAX ADMIN., SIGNIFICANT CHALLENGES EXIST IN

DETERMINING WHETHER TAXPAYERS WITH SCHEDULE C LOSSES ARE ENGAGED IN TAX ABUSE (2007), https://www.treasury.gov/tigta/auditreports/2007reports/200730173fr.html.

68. Id. 69. Id. 70. Over time, uncertainty about whether an activity is a hobby or not is often resolved. An

unending sequence of losses year after year can hardly have an explanation other than that the taxpayer is engaged in the activity other than for profit. It is the requirement of annual tax accounting that compels a premature, or at least only partially informed, determination about the nature of the activity. As a practical matter, taxpayers are only at risk of having their activities classified as hobbies if their expenses exceed their income from those activities.

71. See discussion supra Part I.B. 72. Jim Wang, Does the IRS View Your Side Hustle as a Business or Hobby?, U.S. NEWS (Apr.

1, 2014, 8:58 AM), https://money.usnews.com/money/blogs/my-money/2014/03/31/does-the-irs-view-your-side-hustle-as-a-business-or-hobby.

73. Tony Nitti, Beachbody Coach? Rodan & Fields Consultant? At Tax Time, Beware the Hobby Loss Rules, FORBES (Feb. 25, 2016, 1:57 PM), https://www.forbes.com/sites/anthonynitti/2016/02/25/beachbody-coach-rodan-fields-consultant-at-tax-time-beware-the-hobby-loss-rules/#4f715f61566e.



When you’re a BeachBody Coach, you have to be a product of the product. That makes your workouts, shakes, workout gear, tv to watch your workouts on, gas mileage to buy the stuff, cups to drink shakes from, workout shoes, utilities to run your business, office space, ect. [sic] tax deductible. Am I making sense yet? You don’t get these kind of tax deductions when you work for someone else! As a BeachBody Coach, you can take a monthly personal expense like Shakeology and make it a business expense that is tax deductible.74

The Treasury Regulations promulgated under § 183 include factors to be considered in determining whether an activity was entered into for profit.75 These factors reflect considerations that had been taken into account by courts before the enactment of § 183 in 1969.76 The Treasury Department did not explain why these factors are sensible when they promulgated the regulations, and courts have been stingy with their reasoning as to why they look to the factors that they do.77 To illustrate the application of the screening principle, I consider each of the factors from the Treasury Regulations, in turn, and focus on whether a valid inference can be made from the presence of that factor to a profit motive.78

Manner in which the taxpayer carries on the activity

The first factor from the regulations is the manner in which the activity is conducted.79 For example, if the activity is conducted in too “casual” a manner, this may be inconsistent with how an ordinary business person would act, a fact that is part of the “data of practical human experience.”80 The regulations look to whether the activity is carried on in a “businesslike manner,” meaning that books and records are maintained. Some courts have found that a separate operating name for the business is particularly

74. b4u2, Business Deductions Help, MR. MONEY MUSTACHE (Feb. 4, 2014), https://forum.mrmoneymustache.com/ask-a-mustachian/business-deductions-help/?action=printpage;PHPSESSID=rm8pi740ldqn5a0t8pvo6pt2r0.

75. Treas. Reg. § 1.183-2 (1972). 76. Burns and Groomer, supra note 37, at 198 (“There is general agreement that the nine factors

were extracted from prior case law.”); Michelle B. O’Connor, Primary Profit Objective Test: An Unworkable Standard?, 27 LOY. U. CHI. L.J. 491 (1995).

77. See, e.g., Curran v. Comm’r, 29 T.C.M. (CCH) 696 (1970). 78. I do not discuss the ninth factor in the Treasury Regulations, whether there are elements of

personal pleasure or recreation in the activity. This factor is in some sense at the heart of the entire inquiry. It allows that the presence of a personal motive may indicate that the activity is not engaged in for profit but that the taxpayer need not be exclusively motivated by profit or have an intent to maximize profits.

79. Treas. Reg. § 1.183-2(b)(1) (1972). 80. Bessenyey v. Comm’r, 379 F.2d 252, 257 (2d Cir. 1967).



important.81 The audit technique guide distributed to revenue agents and tax compliance officers also suggests asking about whether the activity has its own bank account or website.82 If the activity is conducted in the way that other profitable activities of a similar nature are conducted, or the taxpayer adopts or abandons operational methods in a manner consistent with an intent to improve profitability, this factor weighs in favor of the activity being engaged in for profit.

There are three problems with these factors. The first problem is the implicit assumption that taxpayers are motivated by either profit or personal enjoyment but not both.83 Both the person who is engaged solely for profit and the person who cares about profit but views the activity mainly as a hobby will have an incentive to conduct the activity in a manner that increases its profitability. The very fact that the hobby-loss rules exist is premised on the fact that there are taxpayers who derive intrinsic utility from certain activities but would also like to get tax deductions for the expenses that they incur. Since deductions are valuable only for the taxes that they save, it is inconsistent to assume that these taxpayers care about dollars saved in the form of tax deductions but not about income from the activity.

The second problem with this factor is that some of these ways of carrying on the activity do not just result in increased profitability but may also yield enjoyment to the hobbyist. One scholar notes that many actions are “consistent with both pecuniary and nonpecuniary motives. Thus, a dedicated cost-conscious hobbyist as well as a determined entrepreneur might keep good records.”84 The hobbyist might also want to adopt the latest innovations and technologies in her hobby solely because they increase the enjoyment of the activity.85 Like record keeping, adopting new

81. Burns and Groomer, supra note 37, at 205 (“In fact, absence of an operating name has been cited as an indication that a profit motive was lacking.”).

82. IRC § 183: Activities Not Engaged in For Profit (ATG), IRS, https://www.irs.gov/pub/irs-utl/irc183activitiesnotengagedinforprofit.pdf (last revised June 2009).

83. As Professor Lederman notes, the dichotomy between tax-avoidance motives and business purpose is a false one: “First, all profit-motivated transactions in a world with taxes are motivated by post-tax profit. Second, many transactions have both tax and non-tax purposes, and it can be hard to separate and quantify them.” Lederman, supra note 3, at 417–18 (footnote call number omitted). Nevertheless, as Justice Harlan wrote,

[f]or income tax purposes Congress has seen fit to regard an individual as having two personalities: “one is [as] a seeker after profit who can deduct the expenses incurred in that search; the other is [as] a creature satisfying his needs as a human and those of his family but who cannot deduct such consumption and related expenditures.”

United States v. Gilmore, 372 U.S. 39, 44 (1962) (second and third alterations in original) (quoting STANLEY S. SURREY & WILLIAM C. WARREN, FEDERAL INCOME TAXATION (1960 ed. 1960)).

84. Allan J. Samansky, Hobby Loss or Deductible Loss: An Intractable Problem, 34 U. FLA. L. REV. 46, 54 (1981).

85. For example, innovations in open source or peer-to-peer sharing software were developed by hobbyists not necessarily seeking profit. See John McGaraghan, A Modern Analytical Framework for



technologies is costly but not costlier for people who aim to make a profit. As such, they are not good screening factors. Thus, even if the hobbyist is not motivated by profit, adopting these ways of conducting the business is not likely to be costlier for her than someone who is engaged in the activity solely for profit.

The expertise of the taxpayer or his advisors

The regulations provide that extensive study of the business, scientific, and economic practices of the activity, or consultation with experts about these matters, may indicate that the taxpayer has a profit motive, particularly if the taxpayer adopts these practices.86 This factor suffers from some of the same problems as the first factor. Hobbyists and those motivated solely by profit alike might want to learn about the business and the economic and scientific aspects of the activity. In fact, avid hobbyists, those who are likely to derive the most enjoyment from an activity, may also be the ones who are most likely to enjoy exploring these other aspects of the activity through study or consultation with experts and other aficionados. This sort of study both generates pleasure for the hobbyist as well as increased profit potential. There is also no reason to think that engaging in this study or consultation with experts will be costlier for individuals who enjoy the activity than those who do not, meaning that they are unlikely to be useful screens.

The time and effort expended by the taxpayer in carrying on the activity, and the taxpayer’s financial status

The regulations suggest that if the taxpayer devotes a considerable amount of time and effort to an activity, particularly if she withdraws from another occupation to do so, then it is more likely that the activity is engaged in for profit.87 This factor disadvantages taxpayers who start a business in their spare time, on weekends and evenings, as compared with taxpayers who may be able to sacrifice income from another occupation to pursue a life of leisure.

The problem with this factor as evidence of profit motive is its reliance on the assumption that an individual who derives enjoyment from an

Monopolization in Innovative Markets for Products with Network Effects, 30 HASTINGS COMM. & ENT. L.J. 179, 187 (2007).

86. Treas. Reg. § 1.183-2(b) (1972). 87. Courts have also looked to the amount of time and labor expended by the taxpayer in

engaging in an activity as proxy for profit motive because the definition of business offered by the Supreme Court in Flint specifically contained a mention of activities that consume “time” and “labor.” Flint v. Stone Tracy Co., 220 U.S. 107 (1911); see Wilson v. Eisner, 282 F. 38 (2d Cir. 1922).



activity cannot also be motivated to earn income from it. As between two taxpayers, both of whom enjoy receiving cash income but one of whom does the activity as a hobby and another who does not, the one who enjoys the activity will spend more time on that activity than the one who is motivated solely by profit. Taken on its own, this factor has perverse implications, screening for the wrong kinds of taxpayers. As I discuss later, however, if the taxpayers’ efforts are interpreted in light of their financial status, more reliable inferences can be drawn.

Expectation that assets used in an activity will appreciate in value

This factor does not helpfully distinguish between individuals who do and do not receive a considerable consumption benefit from the activity because both kinds of taxpayers will be more likely to undertake an activity if there is the expectation of property appreciation. The factor does serve a useful function in making clear that the expectation of income can include all income, including gains from property, as well as cash flows. Of course, if the taxpayer expects that the value of any property used in the activity will lose value, then this factor would be indicative that the activity is a hobby.

Success of the taxpayer in carrying on similar activities

This factor encourages the factfinder to consider that an activity may be engaged in for profit, even if it is presently unprofitable, if the taxpayer has a history of turning around unprofitable activities.88 This factor is helpful in helping courts make predictions about the future profitability of the activity being scrutinized, but it falls largely outside the analysis of this Article because it arises before the taxpayer engages in that activity. Taxpayers are unlikely to attempt to engage in a series of activities with a pattern of early losses and subsequent income solely to create the impression that a later activity will also give rise to income.

The taxpayer’s history of income and losses from the activity and the amount of occasional profits

Finally, the regulations state that when losses are

sustained beyond the period which customarily is necessary to bring the operation to profitable status such continued losses . . . may be indicative that the activity is not being engaged

88. Treas. Reg. § 1.183-2(b) (1972).



in for profit. . . . A series of years in which net income was realized would of course be strong evidence that the activity is engaged in for profit.89

Drawing a connection between realized profits and losses and intentions is tricky,90 but there is a basis for the use of actual income and losses as a factor for inferring intent.

On the one hand, if persistent losses are predictive of future losses, then the presence of persistent losses is a factor that helpfully identifies those who engage in an activity for personal benefit. Indeed, there can hardly be another explanation in such a case since the willingness to tolerate those losses is greater for taxpayers who get a significant consumption benefit from the activity. But the inferences that can be drawn from recurring profits and persistent losses are asymmetric. All individuals presumably enjoy profits, and so the existence of positive profits does not help us discriminate between taxpayers motivated by profit and hobbyists.

Similarly, the pattern of income and losses, whether the venture yields rare but large profits or regular but smaller profits, does not tell us whether the taxpayer is engaged in the activity as a hobby. This factor, like the factor on property appreciation, seems designed to disabuse courts of the notion that unless the income from the activity comes in a particular form at a particular time, then the activity could not be motivated by profit.

Summing up, our application of the screening principle to the factors in the hobby-loss regulations paints a bit of a pessimistic picture. Many of the factors that the Treasury Department has identified as potentially relevant are also those that are easily adopted by people who are motivated primarily by the pleasure of the activity.

Which Factors Satisfy the Screening Principle?

The taxpayer’s financial status rarely plays an explicit or prominent role in hobby-loss analysis, but it should.91 In general, taxpayers prefer to consume a combination of leisure and market goods purchased with cash

89. Treas. Reg. § 1.183-2(b)(6) (1972). 90. Courts also focus on the actual experience with profits and losses as evidence of intent,

primarily because it is objective and because of what they think actual businesspersons would achieve or tolerate in the way of profits and losses. See, e.g., Abrams v. United States, 449 F.2d 662 (2d Cir. 1971); White v. Comm’r, 227 F.2d 779 (6th Cir. 1955); Morton v. Comm’r, 174 F.2d 302 (2d Cir. 1949); Coffey v. Comm’r, 141 F.2d 204 (5th Cir. 1944).

91. Professor Samansky has suggested that wealthier individuals may be more motivated by the pleasure of an activity. Samansky, supra note 84, at 51–52. The legislative history indicates Congress was concerned with wealthy farm hobbyists’ propensity to lose money. See S. REP. NO. 91-522, at 95 (1969) (prior “rules have allowed some high-income taxpayers who carry on limited farming activities as a sideline to obtain a substantial tax loss”).



income. An individual who is wealthier is more likely to reduce the time she spends on her other occupations and use it for leisure than someone who has less wealth. Thus, a taxpayer with a significant amount of wealth who devotes a significant share of her time and efforts to an activity is more likely to be engaged in that activity for personal reasons than someone who is not wealthy but makes the same choices. It is unlikely that the taxpayer will choose to give wealth away merely to facilitate the inference that she is engaged in an activity for profit. Therefore, this factor, in conjunction with the amount of time spent on an activity, satisfies the screening principle.

Having identified the combination of the taxpayer’s wealth and the amount of time that she spends engaged in the putative business as a useful screen for profit motive, it is interesting to compare two examples from the Treasury Regulations under § 183 that illustrate the use of the hobby-loss factors to draw an inference about whether the taxpayers were engaged in an activity for profit. Example 1 and Example 4 both describe taxpayers who inherited, and continued to operate, a nonprofitable farm from a family member.92 The two examples differ in several ways, including that the taxpayer in Example 1 also inherited substantial stock holdings that yield significant dividend income whereas the taxpayer in Example 4 earns only a modest wage as a factory worker.93 The Treasury Regulations conclude that the first taxpayer’s activities, “based on all the facts and circumstances, could be found not to be engaged in for profit,” whereas the second taxpayer’s activities could be found, “based on all the facts and circumstances, to be engaged in by the taxpayer for profit.”94 The tentative conclusions drawn in the regulations, and the likelihood that the outcomes are overdetermined by the number of differences between them, make it impossible to point to the difference in the taxpayers’ wealth and the time they spend engaged in the farming activity as the crucial difference. However, here, as in a number of hobby-loss cases, the taxpayer’s wealth appears to be an important factor.

III. A PRINCIPLED STANDARD FOR FACTS AND CIRCUMSTANCES TESTS

In Part II, I argued that the facts from which the law should infer hidden factors should be those that serve as effective screens for taxpayers with those factors. Specifically, they should be less costly for someone with the factor than someone without it. In this Part, I discuss when the law should incorporate facts that satisfy this screening requirement. That is,

92. Treas. Reg. § 1.183-2 (1972). 93. Id. 94. Id.



should the relevant facts be specified ex ante, by the appropriate regulatory agency, or should the facts be specified ex post, by the courts after observing taxpayers’ conduct?

I analyze this question within the rules–standards framework described by Louis Kaplow,95 but I show that the distinctive nature of laws that depend on private information adds another dimension to Kaplow’s influential analysis. My proposal for how facts and circumstances inquiries should be resolved is a standard, but it is a standard accompanied by a decision criterion for specifying its content.96 Importantly, the choice between a rule and a standard in the context of intent-based tests is a choice about which game is being played between the tax authority and taxpayers.97 Strategic situations in which the regulator specifies the law in advance are known as screening games, while situations in which the regulated party acts first and then the regulator must interpret their actions are known as signaling games. Which game the regulator decides to play has important consequences for the cost of trying to distinguish between taxpayers using facts and circumstances tests.

If the costs of distinguishing between taxpayers are too high, it is more efficient to provide all taxpayers with the same tax treatment than try to induce them to reveal their hidden intentions. Specifically, if a large enough fraction of taxpayers have the tax-favored characteristic, then all taxpayers, both those with and without the characteristic, will be made better off if there is no effort to distinguish between them and instead the difference is split by assigning all of them a tax treatment that is an average of the treatments that the two kinds of taxpayers would receive if their types were known.98 The taxpayers who do not have the tax-favored factor are made better off because they receive better tax treatment than if their intentions were known, and the taxpayers who do have the tax-favored factor are better off because they do not have to incur the costs of signaling to the lawmaker that they have the hidden factor.

In the final analysis, I argue that the law should neither list the facts and circumstances that are relevant in advance, nor should it give courts

95. Kaplow, supra note 6, at 598. 96. Id. at 586. 97. Although the literature on rules and standards is voluminous, the relationship between the

choice between the two and hidden information has been largely unexplored. One exception is Ezra Friedman & Abraham L. Wickelgren, A New Angle on Rules Versus Standards, 16 AM. L. & ECON. REV. 499 (2014).

98. Other scholars have made related observations about the administrative benefits of this sort of “rough justice”. See Daniel Shaviro, The Optimal Relationship Between Taxable Income and Financial Accounting Income: Analysis and a Proposal, 97 GEO. L.J. 423, 437 (2009) (explaining that the 50% deductibility of meals and entertainment can be attributed “to the fact that the recipients of free meals and entertainment, even in a business setting, arguably have economic income that is difficult to tax directly”).



unfettered discretion to identify the factors that are relevant to a mental state inference after the fact. Instead, the statute should specify the principle according to which courts will choose relevant facts and circumstances. That principle is the screening principle.

A. Choosing Between Two Games

In this Part, I show that there is another dimension to the economic analysis of rules and standards that arises when the law depends on private information of the actor: the choice between a rule and a standard is also a choice about whether the regulator is playing a screening game or a signaling game with the regulated parties.

In a screening game, an uninformed party would like to transact with a set of individuals on terms that reflect some trait or characteristic that is private information of those individuals.99 If certain conditions, including the screening principle, are satisfied, then the uninformed party can offer a menu of transactional terms and the informed individuals will select different choices from the menu that reveal whether they possess that characteristic.100 Signaling games, by contrast, share the features of a screening game but differ in that the individuals with private information act first, leaving the uninformed party to interpret their actions.101 Although the difference in who acts first may seem like an innocuous one, it can have important consequences because the potential outcomes of a signaling game may be very different than the outcomes of the screening game.

Consider again the hypothetical case of Tech Corp. and its employee–shareholder Betty from Part II. The lawmaker would like to tax distributions as compensation (deductible by the corporation and included in income at ordinary rates) if they are intended as such, and tax them as dividends otherwise (nondeductible by the corporation but generally taxed to the shareholder at lower rates). The fact that the corporation’s intentions are private information and that corporations may want to pretend that they have a compensatory purpose that they do not have (or deny one that they do) is what makes the relationship between the lawmaker and the corporation a game of asymmetric information. If disproportionate distributions are a strong screen for compensatory intent, then a per se rule that taxes disproportionate distributions as compensation will allow corporations with compensatory intent to make such distributions, but deter corporations from trying to mask dividends as compensation because the

99. See John G. Riley, Silver Signals: Twenty-Five Years of Screening and Signaling, 39 J. ECON. LIT. 432 (2001).

100. Id. 101. Id.



costs of doing so (overcoming shareholder opposition) will be too great. In this way, taxpayers reveal whether they have a compensatory intent or not by their choices. Specifying the tax treatment of disproportionate distributions in advance would make this a screening game.

Alternatively, the legislature could leave unspecified those facts on which compensation treatment will depend. Under those circumstances, taxpayers will structure their dividend and compensation policies with some expectation about how payments will be treated by the courts, including that disproportionate distributions will tend to be interpreted as compensation. A strategic situation like this, in which taxpayers choose how to arrange their affairs and those affairs, including how to tax disproportionate distributions, are interpreted by courts as evidence of their intentions, is a signaling game. Although the only difference between a signaling game and a screening game is whether the lawmaker specifies the tax treatment of the transaction before or after the taxpayer chooses what to do, this difference is important because it introduces uncertainty about how judges will interpret the taxpayers’ conduct.

B. A Principled Standard

Suppose that we want the legal consequence of an action to depend on private information of the actor, and that there is a collection of facts and circumstances that we know satisfy the screening principle. When should the law specify those facts and circumstances? One possibility is to specify those facts and circumstances in the applicable statute or regulations. The alternative is to leave it to courts or to the agency that enforces the law to observe the regulated parties’ decisions and then determine what facts to take into account and how they will be interpreted. This choice, about whether the content of the law is specified before or after the regulated parties have acted, is one of the principal differences between a rule and standard.102 As Louis Kaplow puts it: “the only distinction between rules

102. Kaplow, supra note 6, at 569. This is not quite the same distinction given by Judge Posner and Isaac Ehrlich in their important article on this topic. For them,

[A] standard indicates the kinds of circumstances that are relevant to a decision on legality and is thus open-ended. That is, it is not a list of all the circumstances that might be relevant but is rather the criterion by which particular circumstances presented in a case are judged to be relevant or not. In an automobile collision case governed by the negligence standard these circumstances would be the speed and weight of the vehicles, their design, the time of day, the layout of the highway, the weather, and any other factors that might affect the question how the sum of the expected accident costs and the accident-avoidance costs could have been minimized.

Ehrlich and Posner, supra note 43, at 258. I do not think that this is quite right. Stating the normative criteria is not the same thing as specifying the principle according to which facts will be deemed relevant, and it appears to be simple common sense that is doing the work in this example. It is left to the factfinder to know the empirical relationship between the facts and the normative criterion. I



and standards is the extent to which efforts to give content to the law are undertaken before or after individuals act.”103

Recall that § 183 limits the deductibility of expenses in the case of an activity “not engaged in for profit.”104 This is a standard, because all of the details about which activities, and in which manner conducted, will trigger the application of § 183 are left for the courts or the Treasury Department to work out. However, Treasury Regulation § 1.183-2 provides a nonexhaustive list of factors that are to be taken into account when determining whether an activity was engaged in for profit.105 The factors fill in some of the detail of how § 183 applies and make it more rule like. Because the factors are nonexhaustive, this places the hobby-loss rules somewhere between a rule and a standard. As Kaplow notes, many legal commands occupy this space between fully specifying the conditions under which certain legal consequences attach (a rule) and merely “asking an adjudicator to attach whatever legal consequence seems appropriate in light of whatever norms and facts seem relevant” (the loosest kind of standard).106

There is a voluminous literature on rules and standards, and performing an all-things-considered evaluation of the choice between the two in the context of facts and circumstances determinations is beyond the scope of this Article. However, we can generate some helpful insights by restricting ourselves to the cost-minimization framework provided by Kaplow. This is only one way of evaluating this important choice in legal design, and I make no claims about how the choice would be resolved if one were to take account, for example, of differences in institutional competence between the legislature and the courts, or any other considerations related to the delegation of rulemaking authority that lie outside of this simple cost-minimization framework. My goal in this Part is to explore how the choice between a rule and a standard would be assessed within this framework, and where it might need to be extended.107

propose that factfinders be given more information about which facts are relevant. My approach will still require them to know whether a given fact has a particular screening property. Of course, another important difference between a rule and a standard is who gives content to the law. Frederick Schauer argues that whether the law ultimately settles on a rule or standard may be outside of the control of the rule maker. Frederick Schauer, The Convergence of Rules and Standards, 2003 N.Z.L. REV. 303, 311. If that is the case, then the recommendations in this paper may be moot; however, they would provide a framework for evaluating whether the equilibrium is a desirable one.

103. Kaplow, supra note 6, at 560 (emphasis omitted). 104. Treas. Reg. § 1.183-2 (1972). 105. Id. 106. Kaplow, supra note 6, at 562. 107. There is also much to be said about the interplay between rules and standards, including

whether the latter necessarily evolves towards the former as judicial decisions become precedents for subsequent ones. Although a court’s inclusion of certain facts as relevant in drawing an inference may



In Kaplow’s framework, the normative aim of the law may be public information but the content of the law, that is, the assignment of legal consequences to actions and private information of the actor, is unknown to the government, courts, and regulated parties alike.108 Giving content to a law means refining that assignment process and discovering in more fine-grained detail what should be taken into account in assigning those consequences. This process of investigation is costly.109 Since a rule has more content (i.e., more fine-grained distinctions) than a standard, the cost of promulgating a rule is greater than the cost of promulgating a standard.

Although it is less costly to promulgate a standard than a rule, standards leave actors with less guidance about how to comply with the law. Although those actors can learn about what the standard requires of them by hiring a lawyer and getting legal advice, this is costly. In general, it is costlier to learn about what the law requires if it is formulated as a standard rather than as a rule. Rules are also more likely to induce compliant behavior than standards because actors are more likely to learn about what the law requires and adjust their conduct accordingly.110 Rules and standards also differ in the costs of enforcement; the cost of an enforcement proceeding is greater under a standard than a rule. In Kaplow’s framework, the social objective to be pursued in choosing between a rule and a standard is the minimization of these promulgation, compliance, and enforcement costs.111 Because promulgation costs are incurred once, but enforcement and compliance costs are incurred each time the law applies, rules become more attractive as the frequency with which a law applies increases.

Let’s return to the policy goal of limiting the deductibility of hobby losses, and consider two extremes: the statutory standard and a bright-line rule.

Standard: No deduction is allowable for an activity not engaged in for profit.

be helpful to taxpayers trying to predict what future courts will do, the decisions of factfinders about which facts to include are not legally binding on future decision makers.

108. Kaplow, supra note 6, at 584. Here, it is worth highlighting that the uncertainty about the content of the law is exogenous. For example, the Appendix of Kaplow’s 1992 paper considers the case in which an individual is strictly liable for a harm-producing activity. Id. at 624. Neither the legislature nor the individual nor the courts know what the magnitude of the harm is. Id. Resolving this uncertainty will involve learning facts about the world, such as identifying the negative effects that are caused by the activity and quantifying the harm of those effects. In contrast, the law I am concerned with turns on private information of the taxpayer.

109. Kaplow, supra note 6, at 569 (“[T]he ideal content of the law with respect to these issues [such as the criteria to be used in determining the assignment of liability] is not immediately apparent. Rather, some investigation and deliberation is required.”).

110. Kaplow, supra note 6, at 577. 111. Id. at 625.



Rule: No deduction is allowable for an activity that is conducted under circumstances , , ,... Consider now the costs of promulgation, enforcement, and

noncompliance. The costs of promulgating a standard for hobby losses (existence of a profit motive) are lower than the costs of promulgating a rule that would apply in detail to all circumstances, but the costs of interpreting and enforcing a standard are higher than those for a rule. Is a rule or a standard more likely to be complied with? This depends on how costly it is to the taxpayer to learn what the law requires of her. Thus, in answering this question it is helpful to consider three cases: when the benefit of learning what the law requires is (1) less than the costs of learning about how both the hobby-loss rule and the hobby-loss standard apply, (2) greater than the costs of learning about how both the rule and standard apply, and (3) greater than the costs of learning about how the hobby-loss rule applies but less than the cost of learning how the standard would apply.

The benefit to the taxpayer is the deductibility of her hobby or business expenses. If the benefit to taxpayers of complying with the law is too small to justify the costs of learning about its application under a rule or a standard, taxpayers will conduct their hobbies in the same way regardless of how the law is promulgated. Thus, in the first case, a standard for hobby losses will be preferable to a rule if the additional cost of promulgating a rule is greater than additional costs of enforcing a standard. The promulgation costs of identifying good screening factors across the entire range of hobbies seem prohibitive and therefore it seems likely that a standard would be better in this case.

In the second case, individuals determine that it is in their interest to learn the content of the law regardless of whether it is a rule or a standard, and so rules become more desirable. Although rules are costlier to promulgate, it is less costly for taxpayers to learn how a bright-line rule applies and less costly for courts to enforce. It is of course difficult to quantify the cost to hobbyists of learning about the standard for whether their costs will be deductible, but given the number of hobbyists, these aggregate costs could be quite large indeed. Thus, in this case it is likely that a rule will be preferable to a standard.

In the third case, taxpayers become informed under a rule but not under a standard, which means that their behavior will differ depending on which is enacted. In this case, a standard will be more efficient if the additional promulgation costs of a rule are greater than the net benefit to individuals of acquiring information about the law, plus the difference in enforcement costs. For the relatively low-stakes situations in which hobby losses are implicated, the benefit to taxpayers of compliance is likely to be small, and



the difference in enforcement costs is ambiguous, so in this case a standard is likely to be preferred over a rule.112

One more cost must be taken into account. The facts and circumstances on which tax consequences depend should satisfy the screening principle. But whether a particular activity is likely to have a significant personal element, and whether particular ways of carrying out that activity are likely to be differentially costly depending on how much someone enjoys the activity, may not be known by people who are far removed, geographically, socioeconomically, and culturally, from the taxpayer. The same facts that may be good screening factors in one place may not be good screening factors in another. Implementing the screening principle demands much of the factfinder’s knowledge of the preferences of people engaged in the activity, so it is valuable for the law to preserve the flexibility for factfinders to identify, based on their knowledge of local preferences, the facts that would satisfy the screening principle. This is another reason in favor of a standard for facts and circumstances tests.

Reviewing these three cases, it is unclear whether a rule or standard is, in general, preferable. On the one hand, the costs of promulgating a rule that identifies in advance all the most helpful screening factors across all possible hobbies seem prohibitive. On the other hand, the costs to taxpayers under a standard of discovering how the standard would apply to their circumstances, given all the uncertainty of what factors a court might choose, also seem high.

Perhaps there is another option, with the advantages of a standard but without many of the disadvantages. Consider the following:

Principled Standard: No deduction is allowable for an activity not engaged in for profit, as determined by facts and circumstances that satisfy the screening principle.

This is a standard, because the content of the law (the factors from which intent will be inferred) is not specified in advance.113 It is a principled standard because it also specifies the principle according to which the content of the law will be chosen.114 To be sure, mere standards

112. For an analysis of bargaining in the shadow of a rule or standard, see Jason Scott Johnston, Bargaining Under Rules Versus Standards, 11 J.L. ECON. & ORG. 256 (1995).

113. Strictly speaking, the facts from which a court will draw an inference about the taxpayer’s intentions are not the “law.” From the perspective of the taxpayer who is interested in predicting what legal consequences will follow from her conduct the difference is immaterial; however, it is unquestionably a slight abuse of the term to refer to them as “law.”

114. Posner and Ehrlich assert that “[t]he term ‘principle’ has been used in discussions of judicial decision-making to denote a maxim, sentiment, or policy informing the decisional process.” Ehrlich and Posner, supra note 43, at 259. This is very different than what Ronald Dworkin called a principle. Ronald M. Dworkin, The Model of Rules, 35 U. CHI. L. REV. 14, 22 (1967) (“I call a ‘principle’ a



often have in the background some sort of principle for giving content to the law,115 but the screening principle is not a normative criterion in the sense that the risk-utility standard116 or the Hand rule117 are. The facts that satisfy the Hand rule do constitute negligence and are for that reason normatively significant, whereas facts that satisfy the screening principle are not themselves constitutive of some desirable (or undesirable) state of affairs, but they permit a valid inference to a normatively relevant (but hidden) fact.

The principled standard has some desirable features. First, as a standard, it avoids the potentially large promulgation costs of identifying which factors should be used as screens in the varied circumstances in which taxpayers allege that they are engaged in a hobby for profit. Second, the costs to taxpayers of learning about which facts will be considered are relatively low because of the principle by which they are chosen. Recall that the distinguishing feature of tax laws that depend on intentions is that the content of the law depends on private information of the taxpayer: whether a feature of how the activity is conducted is less costly for them than someone without the tax-favored intention.118 Taxpayers should know whether a particular way of conducting an activity or carrying out a hobby is costlier to them because of the benefits that they derive from the activity. As a result, there is no need for taxpayers to conduct research to predict the content of the law, and the costs of learning about how the standard applies are lower than they would otherwise be.119 Assuming that courts accurately apply the law and interpret the taxpayer’s actions, the taxpayer can, through introspection, reason to the content of the law.

By reducing the cost of learning about the law, more taxpayers will learn what the law requires, with the result that fewer people will be deterred from engaging in side businesses because of the possibility that

standard that is to be observed . . . because it is a requirement of justice or fairness or some other dimension of morality.”).

115. Mark Gergen notes that “[i]deally, a standard states the normative criteria that officials will apply case-by-case to determine the law’s command. The risk-utility standard for design defects in the law of products liability is a good example.” Mark P. Gergen, The Common Knowledge of Tax Abuse, 54 SMU L. REV. 131, 143 (2001).

116. See Ehud Guttel, The (Hidden) Risk of Opportunistic Precautions, 93 VA. L. REV. 1389, 1392 (2007).

117. See id. at 1390. 118. Moreover, when individual information acquisition costs are low, Kaplow argues that self-

reporting is an efficient way to implement complex rules. Louis Kaplow, A Model of the Optimal Complexity of Legal Rules, 11 J.L. ECON. & ORG. 150, 152 (1995). He argues that self-reporting is more efficient when private information costs are much less than the authority’s costs of differentiation. Id. at 157. This is precisely the case here, where an individual knows how her factors will be interpreted. For a related economic analysis of rule “precision,” see Colin S. Diver, The Optimal Precision of Administrative Rules, 93 YALE L.J. 65 (1983).

119. Kaplow is skeptical that learning about the content of the law can ever be easier under a standard than a rule. Kaplow, supra note 6, at 597–98.



they will not be able to deduct their expenses, and more people will be deterred from trying to claim deductions for hobby expenses because they will not succeed. The fact that the content of the law should largely be available through introspection also eliminates the disparities in access to the law that arise when the content of the law is determined through the costly investigation of facts about the world; it will no longer only be those who have access to legal advice who can plan in light of the law.120

Since a principled standard is a standard, it establishes a signaling game between the lawmaker and taxpayers; courts will still be responsible for identifying which facts and circumstances they will use to draw inferences about the taxpayers’ intentions from their conduct. But they will need to choose facts that conform to the screening principle and so their discretion will be limited to consider only those facts that reliably distinguish between taxpayers with and without the tax-favored characteristics. Since taxpayers know that any signals that they try to send to the court about their intentions will be subject to this test, they will not engage in socially wasteful activities that do nothing to credibly convey that they are entitled to the tax treatment they seek. And then, so long as judicial inferences are also subject to the additional guidance discussed in Part C, taxpayers will not feel compelled to engage in unnecessarily costly signaling, and taxpayers can be expected to converge on the most efficient outcome of the game.

C. Rules and Standards as Games

If we choose a rule for facts and circumstances tests, by specifying all of the screening facts in the rules or regulations applicable to the activity, then the tax authority and the taxpayer are playing a screening game.121 In this game, the lawmaker uses her knowledge of differences in the preferences of taxpayers to choose a correspondence between facts and circumstances and a tax treatment; the taxpayers themselves simply

120. I have analyzed facts and circumstances within Kaplow’s framework. That analysis is focused on determining the “ideal” content of the law. This inquiry sets aside the question of how complex the law should be. As Kaplow argues, rules need not be more complex than standards. Kaplow, supra note 6, at 589. Kaplow argues that the question of how complex the law should be is largely independent of the rule/standard choice. Id. But see Weisbach, supra note 35, at 872 (“[T]he optimal complexity of tax rules is greater than the optimal complexity of tax standards.”).

121. A screening game with only one uninformed party is a “monopolistic screening” game. See generally Suren Basov, Monopolistic Screening with Boundedly Rational Consumers, 85 ECON. REC. (SPECIAL ISSUE) S29 (2009). The outcomes of this game differ from those of a screening game in which there are multiple uninformed parties competing in the marketplace. See generally Michael Rothschild & Joseph Stiglitz, Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information, 90 Q.J. ECON. 629 (1976). In the Rothschild and Stiglitz screening game there can only be an equilibrium if the preferences vary sufficiently with type. John G. Riley, Competition with Hidden Knowledge, 93 J. POL. ECON. 958, 964 (1985).



respond by choosing from the alternatives in front of them. The tax authority in this game can induce taxpayers to reveal their intentions by requiring them to conduct their affairs in a way that is differentially costly, depending on those intentions. In order to induce them to separate, the taxpayer with tax-favored intentions must bear some cost that the taxpayer without those intentions is unwilling to bear, even in exchange for better tax treatment.

If we choose a standard for the facts and circumstances test, and leave the courts to interpret the taxpayers’ conduct by applying the screening principle to identify the relevant facts for inferring taxpayer intent, then we also leave it to the taxpayer to choose how she is going to signal to the courts the intentions she had in entering into the transaction at issue.122 The central problem that arises in signaling games is the multiplicity of potential outcomes that arise from uncertainty about how the uninformed party will respond to an unanticipated action by the party with private information. In a signaling game, taxpayers with tax-favored intentions want to send costly signals to the tax authority that they have the tax-favored intentions.

But unlike the screening game, there are a number of possible outcomes of the signaling game in which players fully reveal their intentions, and some of them are much less desirable than others. Which of these “separating” outcomes the taxpayers arrive at depends on how judges would interpret actions that the taxpayers themselves never even take. Those beliefs may cause taxpayers to engage in much more wasteful signaling activity than should be necessary to demonstrate their hidden intentions.

For example, consider again the case of Tech Corp. Suppose that Betty’s work this year resulted in significant revenues from client X, and Tech Corp. would like to issue her a bonus of $80,000 in addition to her pro rata share of any distribution made in respect of her common stock ownership. Tech Corp. may know that a disproportionate distribution to Betty will tend to be interpreted as compensation, but is an additional $80,000 disproportionate enough? If courts believe that only very significantly disproportionate distributions, say, $100,000, are evidence of compensatory intent, then the board may feel compelled to make a $100,000 distribution to Betty to ensure that the payment is deductible by the corporation. Conversely, suppose it has been a profitable year at Tech Corp., and the board of directors would like to issue to Betty a $100,000 dividend in respect of her preferred stock. Doing so will help establish

122. One of the first formal analyses of signaling was given by the economist Michael Spence. Spence, supra note 43. The terminology is borrowed from ROBERT JERVIS, THE LOGIC OF IMAGES IN

INTERNATIONAL RELATIONS 20 (1970).



Tech Corp.’s reputation for making dividend distributions and allow it to sell preferred stock, in the future, at a higher price. However, if the courts believe that any distribution of more than $50,000 is disguised compensation, Tech Corp. may settle for making a $50,000 distribution, notwithstanding its true intentions.

The risk that judges’ beliefs could induce excessive and wasteful signaling suggests that the law should restrict the inferences that judges draw from observing costly signals sent by parties with tax-favored intentions. Specifically, if the relevant law is a standard (including a principled standard) so that the regulator and the taxpayers are playing a signaling game, the Treasury Department should promulgate regulations restricting the inferences that judges can draw in the following way: a taxpayer who incurs some additional cost to signal her intentions shall be assumed to have tax-favored intentions if a taxpayer without those intentions would not incur such a cost, even if she were guaranteed the favorable tax treatment. Adopting this constraint on judicial inferences will ensure that taxpayers are not induced to waste too much effort signaling to the court that they have the relevant tax-favored characteristic when less costly signaling would be credible.123 Returning to the case of Tech Corp., suppose that Andy would accept an $80,000 distribution to Betty as compensation for her services but would not accept a special $80,000 dividend to her, even if that $80,000 would make the payment deductible to Tech Corp. and thereby increase its after-tax income. In that case, it is unreasonable for courts to require anything more than $80,000 as evidence of compensatory intent.

Like the screening game, there are conditions under which it is better not to allow taxpayers to try to distinguish themselves by persuading the factfinder that they have some normatively relevant characteristic. If the share of taxpayers without the characteristic is low enough, or if the cost of sending the signal is not much lower for the taxpayer without the characteristic than the cost to the taxpayer with the characteristic (i.e., if the signal is weak) then the law should not try to distinguish between taxpayers with and without the hidden characteristic, but instead simply assign one tax treatment to everyone that does not depend on private information.

CONCLUSION

I have argued that the law should articulate a principle according to which facts and circumstances inquiries into taxpayers’ intentions should

123. This is an application of the Cho–Kreps “intuitive criterion,” commonly used to restrict equilibria in game theoretic models. See In-Koo Cho & David M. Kreps, Signaling Games and Stable Equilibria, 102 Q.J. ECON. 179 (1987).



be conducted. That principle is the screening principle: the facts from which inferences about intentions should be made should be only those that are much less costly for taxpayers with a tax-favored hidden characteristic than taxpayers without that characteristic. Second, I have argued that considering the role of intentions and other private information adds an important nuance to the debate between rules and standards: the choice between the two is also a choice for the regulator between playing a screening or signaling game with the regulated parties. The differences between these games, including the set of potential outcomes, should be considered when choosing between rules and standards. In many cases, the best way to implement a law may be neither as a rule nor as a standard, but instead as a principled standard.

* * *

1 Hayashi 289-325 - law.ua.edu

Documents

Transcript of 1 Hayashi 289-325 - law.ua.edu